A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 2, 14:45 UTC
Restore `Literal` for `WeightTransferConfig.backend` (#43183)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
May 28, 2026 at 10:39 UTC
2a781756a1a84b2a91294c57dcdc468605cd0feb
1 file changed
+3
-1
vllm/config/weight_transfer.py
+3
-1
@@ -1,5 +1,7 @@
1
# SPDX-License-Identifier: Apache-2.0
2
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
+from typing import Literal
4
+
5
from vllm.config.utils import config
6
7
@@ -7,7 +9,7 @@ from vllm.config.utils import config
9
class WeightTransferConfig:
10
"""Configuration for weight transfer during RL training."""
11
10
- backend: str = "nccl"
12
+ backend: Literal["nccl", "ipc"] | str = "nccl"
13
"""The backend to use for weight transfer. Validated against the
14
`WeightTransferEngineFactory` registry at engine creation time.
15
"""